Bicarbonate activation of the monomeric photosystem II-PsbS/Psb27 complex

Abstract In thylakoid membranes, photosystem II (PSII) monomers from the stromal lamellae contain the subunits PsbS and Psb27 (PSIIm-S/27), while PSII monomers (PSIIm) from granal regions lack these subunits. Here, we have isolated and characterized these 2 types of PSII complexes in tobacco (Nicotiana tabacum). PSIIm-S/27 showed enhanced fluorescence, the near absence of oxygen evolution, and limited and slow electron transfer from QA to QB compared to the near-normal activities in the granal PSIIm. However, when bicarbonate was added to PSIIm-S/27, water splitting and QA to QB electron transfer rates were comparable to those in granal PSIIm. The findings suggest that the binding of PsbS and/or Psb27 inhibits forward electron transfer and lowers the binding affinity for bicarbonate. This can be rationalized in terms of the recently discovered photoprotection role played by bicarbonate binding via the redox tuning of the QA/QA•− couple, which controls the charge recombination route, and this limits chlorophyll triplet-mediated 1O2 formation. These findings suggest that PSIIm-S/27 is an intermediate in the assembly of PSII in which PsbS and/or Psb27 restrict PSII activity while in transit using a bicarbonate-mediated switch and protective mechanism.


Supplemental Text S1
Comparison of the UV-vis absorption and CD spectra (Figs. 2B and 2D) in the two types of PSII monomers, showed only minimal differences limited to the Soret region of the spectra.
Given the lack of carotenoids and chlorophylls in the crystal structures of PsbS (except for an adventitious chlorophyll hydrophobically bound at the interface between the two monomers) and Psb27 (Fan et al., 2015;Xingxing et al., 2018), the presence of these subunits in PSIIm-S/27 cannot be directly responsible for these spectroscopic differences. A more likely explanation is that they arise from the presence of sub-stoichiometric amounts of CP26 and/or CP29 ( Fig. 2A), which were reported earlier as contaminants of PSIIm-S/27 in the mass spectrometry analysis (Haniewicz et al., 2013). These contaminants are also likely responsible for the greater intensity in the steady-state fluorescence emission spectrum at 675 nm (Fig. 2C) and the increased background levels of fluorescence in the kinetic experiments seen in PSIIm-S/27 (see Fig. 3A and below for more details). The increase in the fluorescence could also be due to a specific effect of PsbS and Psb27 binding, resulting in a perturbation of the fluorescence in line with literature reports of similar higher fluorescence background levels in monomeric PSII with bound Psb27 (Regel et al., 2001;Mamedov et al., 2007).

Supplemental Text S2
A large difference was observed when comparing the fluorescence kinetics experiments (Fig.   3A), where both F0 and Fm for PSIIm-S/27 were shifted to higher values compared to PSIIm. This observation could lead to the erroneous conclusion that the presence of the additional subunits caused a decrease in the maximum quantum yield of PSII (Fv/Fm) via a nonphotochemical quenching mechanism (Ruban, 2016). On the contrary, upon close examination of the data presented in figure 3, it seems clear that the two data sets have very similar amplitudes, but with the PSIIm-S/27 curve shifted to higher fluorescence intensities by the addition of a constant fluorescence background. The increased fluorescence does not therefore indicate a change in the quantum yield of charge separation. This is in line with the observed higher steady state fluorescence (Fig. 2C) and is similar to what is seen when comparing cyanobacteria and eukaryotic phototrophs. Cyanobacteria have often been considered to show lower values of Fv/Fm when compared with eukaryotic phototrophs. This has been shown to be due to the phycobilisomes, most probably disconnected from the reaction centers, leading to an anomalous increase in the fluorescence background that results in an apparent reduction of the Fv/Fm (Ogawa and Sonoike, 2016). When this contribution is subtracted, the measured value of Fv/Fm is equivalent to that measured in eukaryotic phototrophs for fully functional PSII (Kalaji et al., 2017;Santabarbara et al., 2019). In a similar way, a constant fluorescence background, which is not contributing to fluorescence changes due to PSII photochemistry, shifts both F0 and Fm in PSIIm-S/27 samples (Fig. 3A).

Supplemental Text S3.
Below is a list of examples from the literature in which small subunits of PSII appear to influence electron transfer through the quinones acceptors of PSII. Some of these might be relevant to the present report.
Mutants lacking PsbJ in tobacco showed an increased lifetime (x100) of the reduced primary quinone QA •and damped oscillations in the flash dependence of the S2QB •recombination, suggesting altered QA to QB electron transfer (Regel et al., 2001). The ΔpsbJ mutant showed an increased level of F0 similar to the observation here in the PSIIm-S/27. Mutants with a disrupted psbX gene showed evidence of reduced binding or turnover of QB (Katoh and Ikeuchi, 2001). Mutants lacking PsbH were found to show slower kinetics of QA •oxidation, but wild-type behaviour was recovered by addition of bicarbonate (Komenda et al., 2002). Both  (Haniewicz et al., 2013;Haniewicz et al., 2015; see Table S1 for details). The bands are numbered as shown in table S1. The lanes labelled as M indicate the molecular weight markers. Figure S2: Fitting of the thermoluminescence data for PSIIm-S/27 in the absence (red circles) and presence (blue triangles) of 5 mM bicarbonate. Measurements were carried out in 20 mM MES pH 6.5, 5 mM MgCl2. A single saturating flash was given at 5 °C and then the sample rapidly cooled to -10 °C. Scan rate was 0.5 °C/s. Fitting was carried out in OriginLab TM using individual gaussians for each component. The black line represents the overall fit. The dark green and cyan lines represent the S2QA •and S2QB •respectively. The orange line represents the contribution of the high temperature band. Supplemental Table S1: Mass spectrometry analysis of PSIIm and PSIIm-S/27. Based on previous results (Haniewicz et al., 2013;Haniewicz et al., 2015), a comparison of the PSII samples ( Fig. 2A and S1) was done with a semi-quantitative approach. As shown, the bands with mass higher than 45 kDa are not directly linked with PSII and appear to be co-purified components. For the PSII components, the main difference between the two samples is the presence of the subunits PsbS and Psb27 in the PSIIm-S/27. We note that PSIIm also seems to lack the small subunits PsbJ, PsbQ and PsbP. We, nevertheless, consider that these differences are due to an artifactually low level of detection of these subunits in the mass spectrometry, because western blot analysis published previously on similar preparations (Haniewicz et al., 2013;Haniewicz et al., 2015), indicated that these subunits were present.